(407d) Industrial Application of a Filtered Two-Fluid Models for Gas-Solid Flows

With the advent of readily-available high-performance computing hardware, the use of computational fluid dynamics (CFD) in the study, design, and scale-up of gas-solid flow reactors and process equipment has become commonplace.  In research-oriented studies, reactor sizes are generally limited to smaller-scale reactors facilitating comparison with experimental data and computational tractability owing to smaller mesh sizes.  In contrast, the industrial practitioner is most interested in simulating industrial-scale equipment either to understand full-scale performance in the context of new process development or to troubleshoot and remediate under-performing production assets.  Agrawal et al (2000) have indicated that the mesh resolution used in simulating rapid gas-solid flows should not exceed ten particle diameters if standard interphase drag laws are used and solution accuracy is to be preserved.  For the industrial practitioner seeking to simulate full-scale process equipment, say a 6 ft diameter x 100 ft tall riser, with FCC-sized granular materials, 750 µm Sauter mean diameter, full CFD analysis poses a formidable challenge.  For this case, simulating even 60 seconds of real time at the required mesh resolutions would require impractically long computing times using today’s state-of-the-art hardware.

Filtered gas-solid flow models can offer relief from the large computational loads required for accurate simulation of commercial-scale equipment.  As discussed by Igci et al (2008), the filtered model seeks to replace the traditional micro-scale drag laws with drag laws better suited for the larger cell sizes required for timely commercial-scale simulations.  The development of such meso-scale drag laws is done by performing a number of simulations in a periodic domain and filtering the results using filters of various sizes.  These filtered results are consolidated into a Richardson-Zaki-type relation that can be used across various cell sizes without compromising accuracy.

In this presentation, industrial experience involving the use of a filtered two-fluid model for simulating gas-solid flows in commercial-scale reactors will be discussed.  The discussion will emphasize hardware modifications conceived with the aid of CFD analysis and, where available, a comparison against post-modification performance obtained in the field.

References

Agrawal, K., P. N. Loezos, M. Syamlal, and S. Sundaresan, “The Role of Meso-Scale Structures in Rapid Gas-Solid Flows”, J. Fluid Mech., 445, 151 (2001). 

Igci, Y., A. T. Andrews, S. Pannala, T. O’Briens and S. Sundaresan, “Filtered two-fluid models for fluidized gas-particle suspensions” AIChE J., 54, 1431-1448 (2008).